NOTE: I originally imagined this post as including embedded Google Trends graphics, which are quite pretty and information-rich. Unfortunately, WordPress is not cooperating with any attempt to embed the code or use plugins to get the graphics, so you’ll have to settle for static pictures and links.
Google Trends is awesome. If you haven’t done so before, it’s a lot of fun (and a semi-productive use of an afternoon) to throw whatever random words come to mind into the Google Trends and watch the interesting patterns emerge. The remainder of this blog post is the result of one such afternoon, spent examining random words, free-association style. The overwhelming conclusion is that the use of words–on the internet at least–is constantly changing. But that’s a much larger point, and outside the scope of the current exercise.
A Positive Control
To get an idea of how Google Trends works, it is instructive to consider a search term with aknown pattern of activity. A starting point for this might be “Guy Fawkes”. Guy Fawkes is connected primarily with the 5th of November, so we might expect that searches for Guy Fawkes would increase yearly in November. What’s more, Guy Fawkes was popularized by the 2006 movie V for Vendetta, whose hero wields Guy Fawke’s mask. V for Vendetta was released in March 2006, so we would expect that there would be a boost in searches around then, as well. Consider the trend:
The expected pattern holds. Interestingly, one can view the expansion of Guy Fawkes geographically in the visualization on the Trends page (under regional interest). Note that it starts exclusively in the UK, moves outward to most English-speaking countries when V for Vendetta is released, and then, for whatever reason, New Zealand decides that Guy Fawkes is awesome and starts searching for him like mad.
I have no idea why New Zealand likes Guy Fawkes so much–preliminary internet searches revealed nothing obvious. Perhaps New Zealanders just have good taste in graphic novels.
The Internet of College Students
I was initially interested in asking about the relative popularity of various academic disciplines through time, but a more interesting pattern began to emerge. Take, for example, this search for chemistry:
You may notice a peculiar oscillation in this graph. Look closely at the months in which chemistry is least and most searched. The troughs fall in months in which school is not in session–December, with its lengthy winter break, and most obviously, the summer. This pattern is not restricted to chemistry; consider computer science:
Same overall pattern, although less distinct. The obvious reason which jumps out is college kids, either researching courses prior to choosing them or in the first few weeks of the class.
Why college and not high school? Consider a class unlikely to be taught at the high school level, political science:
What surprised me about these graphs is just how much they seem to be driven by college students; the seasonal oscillation is by far the dominant pattern. On further reflection, it made more sense to me. Chemists don’t search for “chemistry”, even though they do plenty of searches about chemistry; they search for particular things. Similarly, I suspect that when computer scientists are searching the internet with queries relating to computer science, they don’t necessarily have either of the words “computer” or “science” in the queries. Which fact highlights an important pitfall of Google Trends: there isn’t necessarily any connection between the meaning of a word and the way it is used (e.g. try searching Trends for “tree of life“).
Global warming and seasons
Finally, I wanted to ask about global warming and its pattern of search interest. I hypothesized that global warming, and all the variant words surrounding it (climate change; anthropogenic; greenhouse effect; etc.) would exhibit especially dynamic patterns of search interest, on account of the great political will to relabel or redefine the names for the effect which I refer to for simplicity’s sake as global warming. Witness, for instance, Frank Luntz’s infamous 2002 memo, in which he urges George W. Bush to refer to global warming as “climate change” (one of his focus group participants remarked that climate change “sounds like you’re going from Pittsburgh to Fort Lauderdale”).
Best to plot it overall:
Interestingly, despite Frank Luntz’s efforts, global warming has stayed “global warming”, although “climate change” did surge somewhat during the Bush years.
You might also note something else about the graph, namely that there is another oscillation. Which month has the lowest interest in global warming every year? August. Indeed, interest in global warming wanes in the summer, and paradoxically heats up in the winter.
Initially I was convinced that this pattern emerged from many global warming deniers searching like mad in the winter months, as I have heard the following idiotic argument too many times: “if global warming is happening, why is it so damn cold?” But this appears not too explain it, for if one searches trends for any number of phrases on either side of the global warming “debate”, all terms seem to come with a seasonal oscillation.
What’s more, this oscillation seems to be related specifically to the temperature. Consider this graph of “global warming”, confining ourselves now to Australia:
You can see the oscillation remains, but is inverted, such that the times of lowest interest are in Australia’s summer.
I have no particular interpretation for this finding, and I can’t think of a basis to disentangle all of the possible alternatives. One possibility is that people search most for global warming when it is affecting them (through record high temperatures, for instance), and that such occurrences happen more in the winter than the summer (I believe this is predicted in some climate change models, although now I can’t find a source). Perhaps in the summer, people just take it easy and don’t worry about pressing geopolitical issues less. Conversely, maybe the depressing effect of the winter causes everyone cooped up inside to fret terribly about the fate of the Earth. I’ve no idea.